A word game with ratings that actually measure skill
Why a 1500 in WordSalvo means more than a 1500 in a plain Elo app.
Last updated June 21, 2026 · By Kurt Bijl
Play WordSalvo against real players
What "rated" should mean in a word game
Plenty of word apps show a number next to your name. Far fewer can tell you how trustworthy that number is. WordSalvo runs on Glicko-2, the rating system Professor Mark Glickman of Harvard designed in 2001 as a successor to the Elo system used in chess. The difference matters: Elo gives you one figure and treats every player as equally well-measured. Glicko-2 carries three numbers per player, so a freshly minted 1500 and a battle-tested 1500 are not confused for each other.
In practice that means your rating moves fast and far when the system is still learning you, and settles into smaller, more honest adjustments once it has watched you play. You stop being punished by a single bad night, and you stop being flattered by one lucky bingo.
The three numbers behind your rating
Glicko-2 tracks a rating (R), a rating deviation (RD), and a volatility (σ). Wikipedia's summary of the Glicko system is a good reference, but here is what each one does on your WordSalvo profile.
| Component | What it measures | What you feel |
|---|---|---|
| Rating (R) | Your estimated playing strength on a single scale | The headline number that sorts you into a tier |
| Rating deviation (RD) | How confident the system is in that number — one standard deviation | New accounts swing widely; veterans barely move per game |
| Volatility (σ) | How consistent your results are over time | Streaky players get larger swings; steady players get smaller, calmer ones |
From Novice to Laureate: named tiers
A raw four-digit rating is precise but cold. WordSalvo maps your Glicko-2 rating onto named tiers that run from Novice at the entry end up to Laureate at the top. The tiers give you a goal that reads like progress rather than a spreadsheet cell, while the underlying number keeps the matchmaking honest.
Because RD is part of the calculation, promotion is earned on a body of recent evidence, not one fluke result. Win a string of games against stronger opponents and you climb quickly; the system is confident the jump is real. Beat people far below you and the reward is small, exactly as it should be — Glicko-2, like Elo, weights the opponent's strength into every adjustment.
How WordSalvo compares to other rated word games
Most of the genre still runs on plain Elo or a simple win/loss tally. That is workable, but it has known blind spots: Elo cannot express uncertainty, so a returning player and a brand-new one are treated identically, and a single result can over-correct a number that took months to build.
| Game | Rating model | Tracks confidence (RD)? | Named tiers |
|---|---|---|---|
| WordSalvo | Glicko-2 (rating + deviation + volatility) | Yes | Novice → Laureate |
| Wordfeud | Elo, starting at 1200 | No | Global leaderboard only |
| Scrabble GO | Level / progression-based, not a true skill rating | No | Cosmetic levels |
| Words With Friends | Win/loss record and stats, no public skill rating | No | No |
Why volatility is the part nobody else has
The headline upgrade in Glicko-2 over the original Glicko system is the volatility parameter. Glickman added it to capture how much a player's underlying skill genuinely swings, separate from how confident the rating is. A player who is steadily improving and a player who alternates brilliancies with blunders can sit at the same rating with very different volatility — and the system treats their next game differently because of it.
For a word game that is meaningful. Building a 2,000-rated profile takes consistency across openings, endgames, tile management, and the willingness to fight for every bingo. Volatility quietly rewards the player who is reliable, not just occasionally explosive.
Ratings you can read after the game
A rating only teaches you something if you can see why it moved. WordSalvo pairs the Glicko-2 system with post-game analysis: instant stats plus an engine replay that flags brilliancies, optimal moves, and turning points for covered online and AI games. You learn which decision swung the result, then watch your tier respond over the next few matches.
And the rating is purely earned. WordSalvo has no pay-to-win — spending money on the Ad-Free purchase or the Word Master subscription unlocks themes and analysis, never a gameplay advantage. Your number reflects how you play, full stop.
Frequently asked questions
- what is a word game with ratings?
- A word game with ratings assigns each player a skill number that rises when you win and falls when you lose, weighted by the opponent's strength. WordSalvo uses Glicko-2, a chess-grade system that also tracks how confident it is in your rating and how consistent your results are.
- what is Glicko-2 and how is it different from Elo?
- Glicko-2 is a rating system Professor Mark Glickman published in 2001 to improve on Elo. Where Elo carries a single number, Glicko-2 also tracks a rating deviation (confidence) and a volatility (consistency), so it adjusts faster for new players and more carefully for established ones.
- what are the rating tiers in WordSalvo?
- WordSalvo maps your Glicko-2 rating onto named tiers that run from Novice at the entry level up to Laureate at the top. Climbing a tier requires consistent results against comparable or stronger opponents, not a single lucky game.
- does WordSalvo rate games against AI opponents?
- WordSalvo offers AI opponents at easy, medium, hard, and expert, plus post-game engine analysis for covered AI games. Your competitive Glicko-2 rating is built primarily from matches against real players, which keeps the ladder meaningful.
- can I buy a higher rating in WordSalvo?
- No. WordSalvo has no pay-to-win. Premium purchases unlock themes, ad removal, and post-game analysis, but spending money never affects gameplay outcomes or your rating. Every point is earned at the board.
- why does my rating swing more when I am new?
- New accounts have a high rating deviation — the system is not yet sure where you belong, so it moves your number in large steps. As you play more recent games, the deviation drops and adjustments become smaller and more precise.